New Ekonomistas post (in Swedish). Here is an English translation.

From 1995 to now, Swedish households’ debt relative to their disposable income has almost doubled, from about 90 percent of disposable income to about 170 percent. The Riksbank maintains that the higher debt relative to income has made the households more vulnerable. “[A]s household indebtedness is now very high the households are more vulnerable” is a typical statement. But during the same period, the households’ assets and net worth has approximately doubled relative to disposable income, and is now about 580 and 410 percent, respectively, of disposable income. Households have thus become twice as wealthy relative to disposable income. Have the Swedish households, with now twice as much wealth relative to income, really become more vulnerable, as the Riksbank seems to imply? 

This issue is discussed in my paper ”Resilience, debt, and net worth: Has resilience increased with higher debt-to-income ratios.” To simplify and isolate the issue, I compare two households, household 1, which has a balance sheet similar to Swedish households’ current average debt, assets, and net worth relative to disposable income, and household 2, which has twice as large a balance sheet relative to disposable income (and thus twice as large debt, assets, and net worth relative to disposable income as household 1).

Table 1 and 2 show the two households’ balance sheets. Real assets refer to houses, cooperative flats, and summer homes. Financial assets exclude collective pensions, currently about 1.2 times disposable income. Household 1 has net worth equal to 4.1 times disposable income; household 2 has net worth equal to 8.2 times disposable income.

Table 1. Balance sheet for household 1; assets, debt and net worth relative to disposable income.

Table 2. Balance sheet for household 2; assets, debt and net worth relative to disposable income.

Which household has a larger resilience against different disturbances, household 1 or household 2? Assume that housing prices would fall by almost 20 percent, so real assets would fall by 0.6 times disposable income for household 1 and 1.2 times disposable income for household 2. Then households 1 and 2 get the new balance sheets shown in tables 3 and 4 (boldface denotes entries that differ from the corresponding ones in table 1 and 2).

Table 3. Balance sheet for household 1 after a fall in housing prices; assets, debt, and net worth relative to disposable income.

Table 4. Balance sheet for household 2 after a fall in housing prices; assets, debt, and net worth relative to disposable income.

For household 1, real assets and net worth have fallen to, respectively, 2.6 and 3.5 times disposable income; for household 2, they have fallen to, respectively, 5.2 and 7 times disposable income. Which of the two households have a larger resilience against the fall in housing prices? Is it really household 1, which has lower debt relative to disposable income?

One way to answer this question is to decide which of the two households you would like your own household to be, after the fall in housing prices. If there would be no fall in housing prices, it is pretty obvious that everyone would prefer to be household 2 rather than household 1. But what if a fall in housing prices would occur?

Personally, I would prefer to be household 2 rather than household 1, both before and after the fall in housing prices. It is true that household 2 looses more than household 1 relative to disposable income, 1.2 times disposable income compared to 0.6. But after the fall in housing prices, household 2 still has twice as large net worth as household 1, namely 7 times disposable income compared to 3.5. In other words, after the fall, household 2 is in a better situation than household 1.

To use net worth as a criterion like this means that assets and liabilities are given the same weight. Assume that, as a precaution, one gives double weight to liabilities as to assets. Then one can form a “resilience index,” that equals assets less two times the debt (equivalently, net worth less debt), relative to disposable income. Before the fall in housing prices, the resilience index is 11.6 – 2*3.4 = 4.8 times disposable income for household 2, 2.4 times disposable income for household 1. After the fall, it is 3.6 and 1.8 for household 2 and 1, respectively. According to the resilience index, that is, with double weight on debt, household 2 is still more resilient than household 1.

In my paper I also examine the resilience to a loss in income due to unemployment during a year. The conclusion is that household 2 is more resilient than household 1 for this disturbance, too. I also take into account the household’s “human capital,” the present value of their disposable income (including collective pensions) less necessary minimum household expenses. This is necessary to assess the response of consumption to disturbances, something that is also discussed in the paper and in this previous post. The conclusion is that the effect on consumption of a fall in housing prices is limited. It can be further reduced by a temporary expansionary monetary policy, with a lower policy rate, or with a low policy rate during a longer period.

It thus seems that the wealthier household is clearly more resilient to different disturbances and thus less vulnerable, in contradiction to what the Riksbank seems to imply. It is actually pretty obvious, that twice as wealthy Swedish households, with now twice as large debt, assets, and net worth compared to they had in 1995, can hardly have become more vulnerable. Instead, they are reasonably more resilient to different disturbances. Had their debt been larger at the expense of assets, it would have been different, but not when assets and net worth have increased at least as much as the debt. In order to assess households’ vulnerability, it is as usual not enough just to look at the debt, which is what the Riksbank seems to do. One has to look at the whole balance sheet and at the whole situation of the households.